PASSIONFRUIT PULSE
This was the week the AI search picture got commercial.
At Marketing Live 2026 on May 20, Google announced ads inside AI Mode answers, Conversational Discovery ads and Highlighted Answers, both Gemini-powered, both placed directly in the AI response. The same week, Google confirmed AI Mode passed 1 billion monthly users and published its first real usage report.
Two things stood out once we read past the headlines. First: the new ad layer doesn't let anyone pay to change the organic answer, it rewards brands that already earn organic citation. Second: Google's own usage data is a content brief in disguise. Most people skimmed it. We didn't.
Here's what we're tracking.
Blog of the week
Google put ads inside AI Mode, the mechanics matter more than the launch
Two formats are in testing. Conversational Discovery ads answer a specific question inside AI Mode with creative Gemini builds per query not a static asset you uploaded. Highlighted Answers place a relevant ad directly inside AI Mode's recommendation list. Both carry a "Sponsored" label and a Gemini-written explainer.
The part most coverage skipped: the ad does not change which sources AI Mode cites organically. It sits alongside the answer, labeled, separate. That separation is the load-bearing principle for every conversational ad product in 2026 Google, ChatGPT, and Copilot all keep paid out of the organic answer.
Two implications worth holding onto:
Organic and paid now share one foundation: The structured product data, clean feeds, and entity-consistent content that make a strong Conversational Discovery ad are the same signals that earn organic citation inside AI Mode. If you've done the GEO work, you already have most of what the ad layer needs.
Paid plus organic compounds: A brand cited organically AND running a Highlighted Answer gets two placements in one response. A brand running only the ad gets the sponsored slot while the organic answer recommends a competitor.
The strategic read: the ad layer is an accelerant on earned citation, not a substitute. Because Gemini generates creative per query, the levers shifted structured product data, asset diversity, and landing page quality now matter more than pre-built creative.
Research you should read
The research: We tracked 21 million AI answer, YouTube citations vary 40× by platform
"Should we invest in video for AI search?" is the wrong question. The right one is "which platform does our audience use?" because the answer changes everything.
Over seven months we tracked 21 million AI answers to see how each engine treats YouTube as an evidence source. The spread is not marginal:
Perplexity: YouTube appears in 39.5% of answers, typically 5 videos deep.
Gemini: 29.9% and climbing fast. It went from 3.1% to matching Perplexity in under a year.
ChatGPT: 6.7%. Mostly how-to and comparison queries.
Claude: 0.0%. Across 3.95M cited sources, not one was YouTube.
The findings that should reshape your video strategy:
View count is not a ranking signal. 42.9% of cited videos had under 10,000 views. One channel earned 170 citations from 6 videos averaging 3,055 views each. AI matches content to the query — it doesn't sort by popularity.
10–20 minutes is the citation sweet spot. Under 5 minutes gives AI nothing to extract. Over 60 dilutes the signal. Shorts are structurally excluded — long-form captures 97.2% of all video citations.
Topical depth beats channel size. The top 100 channels hold just 15.6% of all citations. A brand publishing 5 focused videos in its category can match a creator with 100× the subscribers.
Branded comparison queries hit 97% YouTube presence on Perplexity. When AI evaluates brands head-to-head, it builds the case almost entirely from video.
The barrier to entry is far lower than most brands assume. You don't need an audience. You need 10–20-minute structured reviews in your product category, published consistently, titled around what you tested and what you found.
INDUSTRY SIGNAL
Industry signal: Google's first AI Mode usage report is a content brief in disguise
Google published its first real usage data on AI Mode this week, drawn from a year of internal Search data. Carl Hendy was among the first to pull the strategic signal out of it, and it's worth your attention.
The behavioral shifts:
The average AI Mode search is 3× longer than a traditional search. Follow-up queries are growing 40%+ month over month.
More than 1 in 6 AI Mode searches are multimodal — voice, image, or video. Image-based searches are up 40%+ MoM since launch.
"I" is one of the most common first words. People are treating AI Mode like a conversation, not a query box.
But the line that should change your content roadmap: planning-related queries grew 80% faster than overall AI Mode queries over six months. Decision questions starting with "which" grew 40%. Brainstorming queries ("ideas for," "where should I") grew 30% faster than the baseline.
This is the tell. AI Mode users are not searching for definitions — they're searching for decisions. The content that wins this surface isn't "What is X." It's content that helps someone choose between options, plan a path, and weigh trade-offs. Thin definitional content faces a different competition now: a conversational answer to a multi-part question. Comparison content, decision frameworks, and genuine point-of-view pieces are what the fastest-growing query types actually demand.
If your content calendar is still weighted toward "What is" explainers, the usage data just told you to rebalance.
TOOLS
Product of the Week: Are you even in the AI conversation?
Most brands track Google rankings obsessively and have no idea whether AI engines mention them at all. The two aren't the same measurement, and the gap is where brands quietly lose the next two years.
The view we run weekly: Position Tracking for AI Search. It takes the prompts your buyers actually ask and shows, per prompt, your share of voice, sentiment, and position across AI engines.
A pattern from a recent workspace: a brand held 71.4% share of voice on "[brand] SEO agency reviews" strong, position 1, positive sentiment. But on the queries that actually pull in new buyers "best AI SEO agency for high-growth startups," "best programmatic SEO agency" share of voice was 0.0%. Not low. Absent. The brand was winning the queries from people who already knew them and invisible on the queries from people who didn't.
That's the trap of branded-only AI visibility. It feels like winning. It's actually a measurement blind spot. Defending your branded prompts while competitors own your category's non-branded prompts is how you wake up in 2027 wondering where the pipeline went.
You can't win an AI conversation you're not in. And you can't know you're missing from it without tracking the prompts that matter.
The Passionfruit Playbook
One question to audit: For the 5 non-branded queries that bring you the newest buyers, what's your share of voice across ChatGPT, Gemini, and Perplexity? If you only know your branded numbers, you're tracking the comfortable half.
One quick win: Publish one 10–20-minute video review in your core product category this week. Title it around what you tested and what you found — not "tutorial" or "explained." That single asset can capture the video citation slot for an entire query category on Gemini.
One thing to stop doing: Building "What is X" content as your primary AI-search play. Google's own usage data shows planning and decision queries growing 80% faster. Rebalance toward comparison and decision content.
Creator prompt: "What's a real decision my customer is agonizing over right now and what's the comparison or framework piece only I could write to help them make it?"
Until next week,
Passionfruit Team